The Employment Status of Doctoral Recipients - International Journal ...

International Journal of Doctoral Studies

Volume 7, 2012

The Employment Status of Doctoral Recipients: An Exploratory Study in the Netherlands Rens van de Schoot Department of Methods and Statistics, Utrecht University, Utrecht, the Netherlands, and Optentia Research Program, North-West University, Vanderbijlpark, South Africa [email protected]

Mara Yerkes Institute for Social Science Research, University of Queensland. Brisbane, Australia, and Erasmus University Rotterdam, Rotterdam, the Netherlands [email protected]

Hans Sonneveld Netherlands Centre for Graduate and Research Schools, Utrecht University, Utrecht, the Netherlands [email protected]

Abstract Studies of employment often focus on general labour market developments or the employment status of vulnerable groups concentrated at the lower end of the labour market. In contrast, the employment of highly educated individuals, in particular PhD recipients, has received less empirical attention. This article contributes to this area using data from a web survey carried out among respondents at four universities in the Netherlands. Dutch doctoral recipients have an above-average employment rate of 86 per cent. In addition, when looking at variables related to academic and non-academic employment, demographic variables, such as age and children living in the household, as well as publications submitted and accepted, are more closely related to contract type (permanent versus temporary) than factors such as PhD supervision and labour market preparation. Gender is a particularly important variable related to employment status, with male doctoral candidates more likely to be employed outside academia. We conclude with recommendations for PhD candidates, their supervisors and universities. Material published as part of this publication, either on-line or in print, is copyrighted by the Informing Science Institute. Permission to make digital or paper copy of part or all of these works for personal or classroom use is granted without fee provided that the copies are not made or distributed for profit or commercial advantage AND that copies 1) bear this notice in full and 2) give the full citation on the first page. It is permissible to abstract these works so long as credit is given. To copy in all other cases or to republish or to post on a server or to redistribute to lists requires specific permission and payment of a fee. Contact [email protected] to request redistribution permission.

Keywords: employment outcomes, employment rate, gender differences, labour market status, research experience and individual performance, Netherlands, doctoral degrees.

Editor: Eli Cohen

Employment Status of Doctoral Recipients

Introduction In recent years, the labour market status of PhD recipients has been a continued subject of interest in both the Netherlands and abroad. For example, most Western countries are concerned with competing globally with a strong knowledge society, whereby highly-educated workers play a vital role. In the Netherlands, there is a clear focus on improving the knowledge society, including the explicit ambition to be in the top five of knowledge societies globally, which the government feels requires a strengthening of education and an improvement of performance at the top level (Ministry of General Affairs, 2010). Doctoral education is an important factor in this regard, contributing to the development of the nation’s research capacity and generating a highlyeducated workforce. Previous studies in the Netherlands (Hulshof, Verrijt, & Kruijthoff, 1996; Van der Neut & de Jonge, 1993) have focused on the added value of a PhD (in comparison to a master’s degree) with respect to employment outcomes. There has also been attention in the Netherlands for the expectations recent postdoctoral researchers have regarding their career prospects, in particular opportunities to continue their research within academia (Hoffius & Surachno, 2006). Thus far, the study by Hulshof et al. (1996) is the most comprehensive and theoretical study of the situation in the Netherlands. Over the years that followed, no new studies were conducted on the labour market status of Dutch PhD recipients until 2006, when Oost and Sonneveld (2006) reported on the data disclosed by Dutch research and graduate schools about the labour market prospects of their PhD recipients. Again in 2006, and following up on European initiatives, Hersevoort, Rienstra, and Ter Haar (2007) conducted a preliminary exploration of the employment status or graduate destinations of PhD recipients residing in the Netherlands. As a result, research on the labour market status of PhD recipients in the Netherlands is fragmented. The lack of research in this area is in part due to an absence of data. Neither Dutch universities nor the Dutch Ministry of Education, Culture, and Science have a policy of conducting periodic surveys to evaluate the employment status of PhD recipients. This absence is surprising, given the desire to improve the position of the Dutch economy by further developing the knowledge economy. The absence of structural data collection from these groups contrasts sharply with standard practices in the United States, for example, where data on doctorate recipients has regularly been collected since the 1950s. The Survey of Earned Doctorates has been carried out since 1957, which gathers information on all individuals receiving a doctorate from a US-accredited institution. Data collected includes information on socio-economic characteristics, educational trajectories, and post-education plans. More recently, the UK and Australia have conducted surveys on doctoral recipients as well (Manathunga, Pitt, & Critchley, 2009; UK Grad Programme, 2004), showing increased international attention for the employment status of doctoral recipients. In fact, the lack of comparable data on PhD graduates, for example on doctorate holders and employment outcomes, led the Organisation for Economic Cooperation and Development (OECD), the United Nations Educational, Scientific, and Cultural Organisation (UNESCO) and the European Union’s statistical organisation EUROSTAT to advocate an internationally coordinated collection of data about the career paths of PhD recipients in Europe (EUROSTAT / UNESCO / OECD, 2006). Given the lack of comparable data, the majority of studies that do investigate educational or employment outcomes for PhD graduates often focus on specific areas, such as specific fields of study, including economics (Siegfried & Stock, 1999), sociology (Dotzler & Koppel, 1999), history (Sclater, Rudd, Morrison, Picciano, & Nerad, 2008), science and engineering (Lee, Miozzo, & Laredo, 2010) or the biomedical sciences (Knobil, 1996). Another group of studies are focused on specific topics, such as studies on gender (Mastekaasa, 2005; NRC, 2010), academic employment and mobility (J. Enders, 2001; Huisman, de Weert, & Bartelse, 2002; Musselin, 2004) or doctoral training (Bleiklie & Høstaker, 2004; J. Enders & de Weert 2004). Finally, a number of country-specific studies are available, including Finland (Academy of Finland, 2003), the US (Nerad, 2004), France (Dany & Mangematin, 2004) and the UK (UK Grad Programme, 2004).

332

van de Schoot, Yerkes, & Sonneveld

Only recently has more attention been given to creating comprehensive studies of doctoral trajectories (Sadlak, 2004). Until comparable data across all fields of study is available, our knowledge of the employment trajectories of PhD recipients remains limited. We address this lacuna by presenting information on the employment outcomes and major correlates of these outcomes for PhD recipients at four universities in the Netherlands.

Research Questions The current article answers two research questions. First, what is the existing labour market position of doctoral recipients at the time of graduation in the Netherlands? Second, what factors influence the initial labour market position of recent doctoral recipients in the Netherlands? The first question is aimed at increasing our knowledge about the employment status of PhD recipients. An article by Fox and Stephan (2001) finds, for example, that career preferences of PhD recipients vary by gender and field of study. We also know from previous research in the Netherlands that academic employment is a well-trodden career path for many doctoral recipients, with PhD graduates and postdoctoral researchers often preferring to continue their careers in academia (Hoffius & Surachno, 2006; Keijzer & Gordijn, 2000). This preference has been attributed to a high level of job satisfaction, in particular due to flexible working hours, intellectual stimulation, and a high degree of independence in their job. However, both PhD graduates and postgraduate researchers hold negative perceptions about their career opportunities in academia, in part due to the initial period of employment consisting of numerous temporary contracts (Hoffius & Surachno, 2006). To account for possible diversity in employment outcomes, we analyse, on the one hand, whether doctoral recipients go on to academia or to employment in the private sector and, on the other hand, whether graduates are employed in temporary or permanent positions. The second question is focused on increasing our understanding about which factors correlate with these employment outcomes. According to previous studies, several indicators affecting employment outcomes reflect the quality of the PhD program and labour market preparation: integration into the academic community (also internationally; see Bowen & Rudenstine, 1992; Golde, 2000); sound management of the PhD trajectory (by the thesis supervisor, as well as by the PhD candidate) (Berger & de Jonge 2005; Bowen & Rudenstine, 1992; Hockey 1991; Oost & Sonneveld, 2006; Rennie & Brewer, 1987); broad scope of the program (benefiting labour market versatility) (Bowen & Rudenstine, 1992; Hills, Robertson, Walker, Adey, & Nixon, 2003); promotion of academic independence (direction from thesis supervisors, for example) (Bowen & Rudenstine, 1992; Lovitts, 2008; Rennie & Brewer, 1987); and quality of preparation for the labour market (through career information, for example, or support provided for developing future research proposals) (Austin, 2002; Bowen & Rudenstine, 1992; Oost & Sonneveld, 2006). However, because the research on factors related to employment outcomes is relatively limited, we include not only variables aimed at measuring the quality of the PhD program and labour market preparation (such as expectations of the PhD recipients, and PhD supervision and career guidance) but also a wide range of other possible correlates, including demographic variables, PhD status, previous research experience, and individual performance characteristics. The remainder of the article is structured as follows. In the next section, we discuss our data and methodology. In section three, we present the results of our empirical analysis, looking at the correlations between demographic characteristics, the PhD trajectory as well as supervision and labour market preparation with permanent and temporary employment, both inside and outside academia. In the final section, we discuss these results and make some suggestions for future research directions.

333


Data and Methodology Data and Sampling Procedure The data presented here are the results of a web survey carried out among respondents at four universities in the Netherlands (Delft University of Technology, Erasmus University Rotterdam, Utrecht University, and Wageningen University and Research Centre) between February 2008 and June 2009 (see also Yerkes, van de Schoot, & Sonneveld, 2012). These four universities were used because they are representative of the broader variety of universities in the Netherlands, including a younger university with a more limited disciplinary agenda (Rotterdam), a university focused on the agricultural sciences (Wageningen), a more traditional university with a broad disciplinary agenda (Utrecht), and a university focused on the technical sciences (Delft). The respondents for this survey were approached through the Registrar’s office of the university, in charge of organising the doctoral defence following registration for graduation. We then e-mailed respondents, inviting them to take part in the survey. Respondents were approached a maximum of three times, including reminder e-mails. All of the information collected in this survey, including the e-mail addresses gathered at the start of the research, remains confidential. All variables that contained personal information, such as name and address, name of supervisor, or any other personal identification have been removed for purposes of confidentiality. As noted in Yerkes et al. (2012), survey respondents were surveyed at the moment of registering for the defence and therefore may be perceived by some to not yet be doctoral recipients. However, the structure of the Dutch system significantly differs from other countries. PhD candidates are most often employed as full-time researchers at the university during the completion of their doctorate, a distinction discussed further in the next section. Also, the Dutch system lacks ABD status (i.e., all but dissertation) and registering for the defence is only allowed following official approval of the doctoral thesis by the defence committee. Outside of exceptional cases such as fraud, the degree will be conferred following a primarily ceremonial defence. For these reasons, the participants in our study are referred to as doctoral recipients. Huma ni ti es 6% Soci a l Sci ences 16%

Agri cul tura l Sci ences 7%

Medi ca l a nd Hea l th Sci ences 25%

Na tura l Sci ences 31%

Engi neeri ng a nd technol ogy 15%

Figure 1. Percentages of field of study The sampling frame consisted of 1113 PhD candidates who registered for their defense between February 2008 and June 2009 at the four universities studied. Just over half of those candidates approached agreed to participate in the study and completed the survey. The overall response rate was 50.7% with a survey sample of 565 respondents. The survey sample consists of 54% men,

334


46% women and includes respondents from all disciplines, see Figure 1. The mean age of our respondents was 34 years old with the majority between the ages of 25 and 40. Some graduates were over the age of 40 when reaching the completion stage of their doctorate: Two-thirds of doctoral recipients surveyed were born in the Netherlands (67%). Doctoral recipients born in other countries were most often born in other Western European countries, Asian countries, or Eastern European countries. Less than five per cent of doctoral recipients surveyed were from North America, Latin and South America or Africa.

Measures In the analyses to follow, individual level differences are accounted for using demographic variables, variation in PhD trajectories and the effects of PhD supervision and labour market preparation.

Demographic variables The following demographic variables are included: marital status, the presence of children in the household, nationality/citizenship, gender, and age. Marital status is measured with a dummy variable: the reference category is never married/divorced/widowed/separated; the category of married or cohabitating=1. We also estimated the effect of the presence of children in the household (no children present is the reference category). Given the complexity of nationality, we used three different variables here, including whether or not an individual came to the Netherlands to obtain their PhD or for some other reason, whether an individual has Dutch citizenship or not, and whether or not the individual is living in the Netherlands at the time of the defence. The first variable measured whether or not an individual came to the Netherlands to obtain their PhD. The citizenship variable measured whether or not an individual has a Dutch passport. Living in the Netherlands at the time of the defence is a dummy variable (a country other than the Netherlands is the reference category). Gender is measured with a dummy variable (female is the reference category). Lastly, we measured age in years.

PhD status In the Netherlands it is possible to differentiate between three different types of PhD status (see also Yerkes et al., 2012), including (a) a PhD candidate employed by the university, (b) scholarship recipients, and (c) external and/or dual PhD candidates. Full employment contracts for PhD candidates are the exception and not the rule throughout Europe. Only the Netherlands, Finland, and Turkey have doctoral educational structures in which different types of PhD status exist simultaneously (EC, 2007). In the Netherlands, PhD candidates with the status of employee are covered by an employment contract, which specifies working conditions and salary based on a collective agreement covering the Association of Universities in the Netherlands (see www.vsnu.nl for more information). Scholarship recipients do not have an employment contract. Rather, they are given a scholarship or stipend for a specified period of time. Lastly, external and/or dual PhD candidates, often not accounted for in studies on doctoral education in the Netherlands, are an amalgamation. These PhD candidates do not have a formal PhD contract at the university, nor do they have a scholarship or stipend. Their status can take on different forms, for example an external candidate who works part-time on his or her PhD thesis while having a job elsewhere, or dual candidates, such as junior lecturers who work part-time at a university while working part-time on their PhD thesis. There are also a number of external candidates who work on their PhD thesis during retirement. Throughout this article, these three forms are referred to as (1) employees, (2) scholarship recipients, and (3) external candidates. The majority of respondents surveyed (71.1%) reported that their main formal status was ‘employee’ with five per cent listing ‘scholarship recipient’ as their main PhD status. The share of external or dual PhD candi-

335


dates was 23.9 per cent. The number of scholarship students in our sample is too small to be included in the analyses here. Therefore, we distinguished between PhD candidates employed by the university (the reference category) and external candidates. PhD status was included as a control variable in our analyses.

Previous research experience Research experience was measured as whether or not an individual gained research experience prior to commencing the doctoral trajectory and, if so, the number of years of research experience. We note that the variable measuring years of research experience did not have a standard distribution and was therefore included in the model as a count variable with a Poisson distribution. In other words, nearly all respondents have a zero (no years of research experience) yet the answers provided by respondents who do not have a zero are distributed normally. Mplus accounts for this zero-inflated variable, using a mixture approach in the analyses, which results in two parameters: (1) an intercept and mean score for no experience versus experience and (2) a regression coefficient for the non-zeros.

Individual performance characteristics These were measured in terms of publications, expectations in regards to publications, and entrepreneurship. Publication variables included the number of submitted and accepted international, scientific journal articles. We also included (1) whether or not respondents had individual expectations of how much scientific output they expect to produce (i.e., the number of papers/book chapters), (2) how much scientific output their supervisor expected them to produce, and (3) how much scientific output their research school or institute expected them to produce.

Effect of PhD supervision and career guidance In the survey, we asked PhD candidates to respond to a series of statements regarding PhD supervision and career guidance. We did this based on the assumption that the quality of the PhD program (including supervision) and labour market preparation could be related to educational and employment outcomes of PhD recipients. Respondents were presented with 29 statements related to the PhD trajectory and labour market preparation. Responses were measured on a five-point Likert scale, varying from 1 “completely agree” to 5 “completely disagree.” The lower the average score, the more respondents agreed with the statement. These statements were combined to form nine scales measuring (1) the role of the supervisor in creating the PhD candidate’s network, (2) the role of the supervisor in supporting and preparing the PhD candidate for the labour market, (3) the PhD candidate’s insight into the necessary steps to be taken during the research trajectory, (4) the quality of supervisory guidance in writing and finishing the PhD thesis, (5) the versatility of the educational trajectory (in terms of subjects studied and extra study and research possibilities) and labour market preparation, (6) the intensity of contact with other PhD candidates (preventing isolation during the PhD trajectory), (7) the quality of preparatory labour market information provided for by the supervisor/graduate school/university, (8) individual responsibility of the PhD candidate in finding a job following graduation, and (9) research experience abroad and support in obtaining international research funding post-PhD. These scales were created using confirmatory factor analysis, using Mplus 5.21 (Muthén & Muthén, 2007). Full maximum likelihood estimation was used to deal with missing items. Fit indices indicated a moderate fit (CFI = .90; TLI = .88; RMSEA = .04; SRMR = .06; n = 413).

336


Analysis The analysis consists of two parts. The first research question is answered using descriptive statistics. Note that in the current paper, all available information about our respondents is used. No data have been imputed through missing analysis for the descriptive information, which explains why the sample size does not equal 565 for some of the results. The analyses presented for the second research question are based on logistic regression models carried out in Mplus v.5.21 (Muthén & Muthén, 2007). We included respondents with missing data in the model estimations using full information maximum likelihood estimation because missing data were completely at random (C. K. Enders & Bandalos, 2001). We began the analyses by looking at having a permanent or temporary contract, broken down into academic and nonacademic employment. The analyses used to answer the second research question consisted of three analytical steps: (1) the relationship between demographic characteristics and employment status, (2) the relationship between previous research experience, academic performance, PhD status and employment status, and (3) the relationship between PhD supervision, labour market preparation, the quality of the educational trajectory, and employment status. Note that the latter model is a hybrid model, where the latent variables of the measurement model of the questionnaire are used to predict employment status. The results of our analyses are reported in odds ratios. We only report the odds ratios and significance values for space reasons. The full results, including standard errors and p-values can be obtained from the authors upon request. A simple way to interpret the findings in the tables is to consider whether the odds ratio differs from one, not zero. For example, if we compare men and women and the likelihood of having a permanent contract, an odds ratio of 1 would mean there are no differences between men and women for the likelihood of having a permanent contract, whereas an odds ratio of three would mean that men are three times more likely than women to have a permanent contract. Odds ratios of less than one, for example, an odds ratio of 0.3, would mean women are 1/0.3 times more likely to have a permanent contract than men. Statistical significance is reported as follows: *=p-value